The invention relates generally to a method for automatically recording and replaying video surveillance imagery from multiple cameras. In particular, this invention relates to continuous buffered recording of multiple video streams to capture pre- and post-trigger imagery associated with multiple triggering events, especially for use in a road-mobile vehicle.
Video surveillance to provide situational awareness in combat vehicles demands efficient use of limited computing and networking resources. Triggering events are typically isolated and demand that the surveillance camera already be looking in the direction of the event. Even then, the triggering event causes the capture of images generated after the event, resulting in the loss of actual event data.
Many modem cameras address this problem by buffering the image stream in the camera memory. This limited buffer can then be delivered upon demand through the network. Challenges arise in obtaining video segments that exceed the capacity of the camera's built-in buffer. Additionally, this approach requires that commands be sent to the camera to initiate the downloading of the buffered data. In some camera models, new imagery cannot be buffered until the data transfer has been completed.
Cameras mounted on pan-tilt platforms require response time for the platform to correctly position the camera for aiming at the region of interest and thereby miss the triggering event unless the platform serendipitously points in that direction beforehand.